The King's Chapel at Restalrig and St Triduana's Aisle

7/27/2019 The King's Chapel at Restalrig and St Triduana's Aisle

1/26

THE KING'S CHAPEL AT RESTALRIG AND ST TRIDUANA'SAISLE: A HEXAGONAL TWO-STORIED CHAPEL

OF THE FIFTEENTH CENTURYby IAIN MAcIVOR, B.A., F . S . A . S G O T .

TH E sunken vaulted chamber of a fifteenth-century hexagonal building adjoins thepartly rebuilt medieval church of Restalrig. Only fragments remain of an upperchamber of the hexagon. The lower chamber, used as the laird's burial vault in theseventeenth and eighteenth centuries, was cleared out and restored in 1907 by DrThomas Ross, and has since been known as 'St Triduana's Well-House'. In 1952the Church of Scotland Trustees gave the hexagon into the guardianship of the thenMinistry of Works. Before repairs could be carried out it was necessary to regulatethe water-level in the sunken and flooded interior. An investigation begun to estab-lish the original water-level showed that the presence of water was accidental, andsuggested that the lower chamber was not a well-house but a chapel, part ofJamesIll's capella regis. This paper describes the hexagon, examines its medieval andmodern history, and discusses the affinities and inspiration for an apparently uniquedesign.

I am especially indebted to my colleague Mr John Cartwright, who supervisedthe survey and drawing throughout, and to Mr Christopher Connell, who is respon-sible for the drawings as they appear in this paper, and Mr Angus Fraser, whoexecuted the first set of drawings from a survey made in 1957. I wish also to thankmy colleagues Mr Thomas Dalrymple, Mr George Hay and Mr James Pugh fortheir help, advice and co-operation. And I gratefully acknowledge the assistanceI have received on continental material from Mile A. Neury, Centre de recherchessur les monuments historiques, Paris, Mr C. J. A. C. Pieters, Rijksdienst voor deMonumentenzorg, the Hague, and Dr W. Zimmermann, KunstdenkmaleraufhahmeRheinland, Bonn.

I. D ESCRI PTI O NThe ancient village of Restalrig (G.R.: NT 284745) now lies within the City of

Edinburgh. A few mutilated houses of the seventeenth and eighteenth centuriesremain. Its rural setting, admired in the early nineteenth century, has been replacedby housing and the St Margaret's motive power depot of British Railways. Thelatter overlies the site of St Margaret's Well, from which the small hexagonalsuperstructure was largely removed to the Queen's Park in 1859. The parish church,hexagon and churchyard of Restalrig lie a short distance to the N. of depot andwell-site.

Restalrig parish church is a rectangular building of four bays measuring exter-nally 66 ft. E. and W. by 23 ft. N. and S. (figs, i and 2). It is a restoration byWilliam Burn in 1836 of the ruins of a fifteenth-century church. Burn erected the


2/26

248 P R O C E E DI N G S OF THE SOCIETY, 1963-64W. gable wall and the western ends of the N. and S. walls. But medieval masonryat the foot of the SW. church buttress and the infilling of the awkward angle betweenthat buttress and the hexagon shows that the W. gable and buttresses were re-erected on the line of medieval structures, which by 1836 must have been in a veryadvanced state of decay. Thus the present church retains the plan of the original.The rest of the present church is largely medieval with 1836 tracery replacementand fabric repairs: drawings made by James Skene in iSiS1 show the E. end andmost of the N. and S. walls standing almost to the wall-heads.

Burn's restoration of the external elevations was faithful. The surrounds andtracery of the three S. windows are competent fabric replacement of originals whichsurvive in detached fragments. In the bay E. of the hexagon, the S. wall has ablocked round-headed door decorated with small tablet flowers. B uttresses at theangles of the E. wall and on the S. wall are medieval. The tracery of the E. windowis probably an authentic copy. B u r n slapped three lights through the N. wall,hitherto without fenestration. The interior of the church is now plastered andcovered by a timber barrel ceiling: no ancient features are visible.

Excavation N. of the church in 1962 revealed a fragment of the footings for aN. sacristy, adjoining the church and contemporary with it. One of Skene's drawingsshows the low roof-raggle of the sacristy, enclosing a blocked doorway. Thesefeatures were visible until 1962. The original northward extent of the sacristy cannotbe defined. The remains of the sacristy were hidden in 1962 by the construction ofa new vestry.

Excavation in 1961-2 also showed the base of a respond, the plinth of whichadjoined the NE. angle buttress of the present church. The respond, with the frag-mentary structural remains around it, seemed to be the eastern termination for anaisle on the N. side of a nave of about the same width as the present church. Therespond was not set out on the same axes as the present church. The mouldings ofthe respond were similar to, but not identical with, the mouldings of the base of thecentral pier of the lower hexagon, suggesting a similar date in the second half of thefifteenth century.

The relationship of the respond to the NE. angle buttress showed that the respondwas constructed after the buttress, the foundation of which had been cut away tomake room for it. As the buttresses are contemporary with the present church therespond is therefore later than the church. (Although the buttresses are of ashlar andthe ancient part of the church is of rubble, the footings show that they are con-temporary.)The only other architectural feature without any obvious structural relevance toeither the present church or the hexagon is the base of a small fragment of innerwall-face, projecting from the external face of the NE. wall of the hexagon. Theshaft-base has the same moulding as the upper part of the respond found by exca-vation and described above. The shaft-base and fragment of wall-face are workedfrom one stone, which is set into the wall of the hexagon. The ashlar face of the1Skene, James, Old Ch a p e l at Restlerig, Edinburgh City Library PYDA/24O6/227I, reproduced as PLXXXVIII; compare PI. XXXIX, I.


3/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA's AISLE 249hexagon is roughly cut away to receive it. The moulding of the shaft-base is stoppedon the oblique line at which the hexagon's wall-face cuts across the stone. Itsrelationship to the hexagon wall shows that it is inserted into the hexagon.

The resemblance of the shaft-base to the respond at the NW. angle of the presentchurch, and the incomplete nature of both, suggest that the two belong to a fabriclater than the hexagon and the present church, a fabric ambitiously begun (thoughperhaps inaccurately set out) and abandoned in the earliest stages of construction.The shaft-base occupies the position of the outer wall of a S. aisle for this proposedbuilding. The latter was designed to have a nave and aisles of unknown westwardextent and a chancel of unknown plan.

The hexagon adjoins the westernmost bay of the present church (figs, i and 2;PI. XXXIX, 2). The lower chamber, its floor 8 ft. below present ground level, isintact. The chamber measures internally 35 ft. across between its opposite angleswithin walls all 3 ft. 6 in. thick except the N. wall which is only i ft. 9 in. thick. TheS., SE. and NE. walls are pierced by windows of three pointed and cusped lightsbeneath three-centred arches (an arch form very uncommon in Scotland) externallysurmounted by moulded dripstones. The splayed outer sills of the windows open topresent ground level. The lower chamber is entered from the churchyard througha doorway set obliquely in the SW. wall, by a flight of steps which continues beyondthe wall faces upwards into the churchyard and downwards into the floor area ofthe lower chamber. The floor of the church is 8 ft. 6 in. higher than the 1907 pavedfloor of the lower chamber.

There is only one original buttress, at the NE. Three additional buttresses werebuilt in 1907. Two angles, obstructed by burial enclosures in 1907, have no buttresses(see p. 256 below). These two angles, at the W. and SW., are finished by a broadashlar chamfer. A near-straight joint runs down most of the height of the buildingat its W. angle, at the arris of the chamfer,1 and contrasts with well-contrivedjointingelsewhere. The chamfers may have been formed and, where necessary, newly faced,in the course of a tidy removal of original buttresses. The wall-heads are finished witha parapet within which rises a pyramidal timber slated roof surmounted by awooden statue of St Triduana: parapet and roof are parts of Dr Ross's 1907restoration.Inside the lower chamber a vault springs from the foliated capitals of angle-shaftsand a clustered central pillar (PI. XL). The usual method by which a polygon withcentral pier is vaulted is by throwing the principal ribs in semicircular or pointedarcs from central pier to wall-shafts. This was the vault-plan of the upper chamber(fig. 36) of the Restalrig hexagon. In the lower chamber (fig. 30, Pis. XLI, i, XLII)the principal arcs are thrown from the wall-shafts at each angle to the wall-shaft atthe angle next but one, while the mid- or ridge-rib first runs horizontally to anintermediate boss and thence curves downwards to the capital of the central pier.

1The joint was revealed in June 1964 when the SW. burial enclosure was demolished. The burialenclosure appears on the photographs, but it is omitted from the plan (fig. i) . A similar joint was observedby Dr Ross in 1907.


4/26

250 PROCEEDINGS OF THE SOCIETY, 1963-64A large recess in the north wall is 11 ft. 3 in. wide and 3 ft. 6 in. deep under a

segmental arch. There is a small aumbry in the NE. wall of the lower chamber. Thelower part of this aumbry was destroyed in the 1907 restoration.Though little survives in position of the upper chamber, there is enough toestablish the main points of its internal design when taken in conjunction with themany associated fragments. A few floor-slabs in position give a floor-level 8 ft. 6 in.above the floor of the church, from which the floors of the two chambers werevertically equidistant. Walling exists in some places to about 18 in. above the floorlevel. The NE. wall is faced with fine ashlar slabs, and in the middle of this wall isthe moulded sill of a recess, either an aumbry or a sacrament-house (PI. XLI, 2).Its position corresponds to that of the aumbry in the lower chamber. In the E. angleis a mass of mortared rubble, presumably the substructure for an altar.

On the floor a large number of dressed stones were discovered when a mound ofearth over the lower chamber was cleared in 1907. These included fragments ofwindow tracery, many rib-stones from the upper vault and parts of the tas-de-charge,window ingoes, and six large ornamental bosses.1 The rib-stubs of the bosses showthat the upper vault had no central pier; its ribs sprang from angle-shafts or corbelsto a central crown (the boss of the crown, or a ring substituted for it, was not found).Though no evidence was discovered to establish the springing-level of the vault, itmay be inferred from the vault construction alone that the upper chamber was muchmore spacious than the lower.There were probably large windows on the three S.-facing walls of the hexagon,corresponding to the fenestration of the lower chamber.

No ancient walling survives on the'N. side next the church, where it is all 1907restoration, so there is no way to determine how the upper chamber was enteredfrom the church.The NE. buttress of the hexagon (the only original buttress) is stopped obliquelyagainst a buttress of the church in a manner which shows that the hexagon is laterthan the church.A second hexagon used to stand a little way to the S. of Restalrig church, within

what is now St Margaret's locomotive depot. This was the little vaulted house ofSt Margaret's Well, the interior of which was dismantled and re-erected in theQueen's Park in 1859, 'happily rescued from the inexorable utilitarianism of themodern railway vandals'.2 The removal was carried out under the direction of thethen Office of Works, following a recommendation by the Society of Antiquaries ofScotland.3 The interior is a miniature parallel to the lower chamber of the Restalrig1The bosses may be briefly described as follows: (i) The lower surface of the boss is finished as a flatdisc bearing, in relief, the letters ' I M S ' surmounted by a crown (PI. XLIII, i); (2) On a disc similar to (i) asaltire flory ( P I . XLIII, 2 ) ; (3) The lower surface of the boss is finished as a shield bearing, in relief, theletter V supported by birds and surmounted by a crown (PI. XLIII, 2); (4) In high relief carved with foliage,the rounded under-surface bearing a mullet (PI. XLIII, 4); (5) The whole surface carved with stylised floriatedornament (PI. XLIII, 5); (6) A grotesque head with prominent upper teeth (PI. XLIII, 6).2 Wilson, Sir Daniel, Memoria ls of Ed inburgh in the olden t ime, 3rd ed. (1891), n, 250.3 Laing, David, 'Statement relative to the removal of St Margaret's Well', in P.S.A.S., in (1858-60),

365-7-


5/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA's AISLE 25!hexagon, with a vault identical in layout but with an internal area less than one-tenth of the interior of the Restalrig building (fig. 30, d) .l

II. THE L E G E N D O F S T T R I D U A N AThe chief source of the legend of St Triduana is the Aber deen Bre v iary , followed by

Constantin Suyskens in the Acta Sanctorum? The substance of the legend is that StTriduana was one of several virgins who accompanied St Regulus to Scotland. Shesettled for a time at Rescobie in Angus, where she was sought after by Nechtan, thelocal ruler. In response to an unhappy compliment by Nechtan on the beauty of hereyes, the saint plucked them out and sent them to him impaled on a thorn. Nechtan'sattentions ceased, and Triduana removed to Restalrig, where she spent the rest ofher life and where she was buried. The Aberdeen Breviary notes that her bones lay atRestalrig and were greatly honoured.

The Acta Sanctorum describes a miracle wrought by St Triduana in which thesaint appears in a dream to a blind Englishwoman, whom she tells to go to her tombat Restalrig. The woman does so and recovers her sight. Later the woman'sdaughter loses her eyes in a fall; her mother prays to the saint and her child's eyesare restored by virtue of the prayer alone, without pilgrimage to Restalrig.

David Chambers gives another version of the legend in which Restalrig is notmentioned. Chambers cites 'Hunter the Dominican' as authority for the widespreadhonour in which St Triduana was once held in Scotland, and for the several shrinesdedicated to her.3 But Hunter's work does not now appear to be extant: at leastSuyskens could not find it.4The repute of St Triduana continued up to the Reformation. Sir David Lyndsaymentions those who went to the saint 'to mend thare eine'.5 This presumably refersto Restalrig.6

There is no reference in any of these writings on St Triduana which suggests awell or spring connected with the saint.

III. M E D I E V A L HISTORYSuyskens thought that the traditional sixth century date for St Triduana might

with luck be correct, but was unconvinced.7 It is not unlikely that there was achurch at Restalrig in the Dark Ages, though there is no evidence for one. Rectors

1The Royal Commission on the Ancient Monuments of Scotland, An Inventory of... the City of Edinburgh(1951), 239.2 Breviarium Aberdonense, Bannatyne Club (1854.), Pars Estiva, October, fol. cxxii. Acta Sanctorum, Octobris,iv, Brussels (1856), 280.3 Chambers, David (Camerarius, Davidus), De Scotorum Forlitudine, Paris (1631), 179: 'Magno fuisse inhonore hanc sanctissimam Virginem apud Scotos ob vitae sanctitatem & miraculorum gloriam testaturHunterus Dominicanus, sacellaque varia in eiusdem honorem extructa fidem faciunt.'4 Acta Sanctorum, loc. cit.6 Th e Monarc he , 1. 2366.6 There are St Triduana place-names at Kintradwell in Sutherland S. of Helmsdale and a St Tredwell'schapel and St Tredwell's Loch at Papa Westray in Orkney: R.C.A.M., A n Inventory of Orkney (1946), N o. 521.The healing of the stabbed eyes of John Bishop of Caithness, c. 1201 at the tomb of the saint, mentionedin the O rkney inga S a g a (ed. A. B. Taylor (1938), 346), may have taken place at one of these northern shrinesrather than at the Restalrig tomb. In the saga the saint's name is Trollhcena.' Acta Sanc torum , loc. cit.


6/26

252 PROCEEDINGS OF THE SOCIETY, 1963 64or vicars of the church are recorded from the twelfth century.1 But our immediateconcern with the church does not begin until the end of the middle ages.

A charter dated I4772 states that James III endowed a chaplainry in the parishchurch of Restalrig for the altar founded by the king in the upper chapel, in cape l lasuperiore- of the same church. The dedication of altar and chapel are not given. Anentry in the Exchequer Rolls for 1486-7 records payment to a mason for stoneroofing-slabs for the King's Chapel adjoining the church of Restalrig.3

When the parish church of Restalrig was erected into a collegiate church byJames III, a papal bull of Innocent VIII dated 1487 stated that James had built a'sumptuous new work' at his own expense.4 It has been inferred from this bull thatRestalrig Church was rebuilt in its present form in I487.5 But in the light of theExchequer Rolls entry quoted above, which refers to work on the roof of a buildingcalled the King's Chapel in the year previous to the issue of the bull, it seems morereasonable to identify the 'sumptuous new work' of the bull with the King's Chapeladjacent to the church.In 1497-1507 John Fraser, Bishop of Ross and sometime dean of Restalrig,founded the chaplainry of St Triduana in St Triduana's aisle (insula) in the collegiatechurch of Restalrig. This chaplainry was erected by James IV in 1512 into theeighth prebend of the college.6

The 1515 charter of the college lists the perquisites and responsibilities of thisprebendary: 'The eighth prebendary who will be entitled of St Triduana's aisle willhave for his sustenance ... (a list follows) . . . and also all the oblations made to thealtar and the relics of St Triduana in connection with the chamber and garden nearthe college. In consideration of which he will sustain and maintain the lower aisleof the said college and the altar of St Triduana situated therein in ornaments, books,chalices, wax candles, and other necessaries in all honesty at his own charges.'

This lower aisle, dedicated to St Triduana, is situated near (prope) the college.The King's Chapel has been described as situated adjoining (contigua) the church.The 1477 charter mentions an upper chapel with an altar endowed by James III.It seems reasonable to bring together the upper chapel and the lower aisle as elementsof a two-storied building known as the King's Chapel, the sumptuous new workreferred to in Innocent VIII's bull. There is no opposition between capella and insula,used of the upper and lower chambers respectively. Capella is a term of status, insulaa term of structure. The most obvious use of inferior and superior is to express verticalsuperimposition, and ecclesia inferior is so used of the lower church of GlasgowCathedral.7The Restalrig hexagon adjoins the church, it had two storeys. Thus the evidencesuggests that the hexagon isJames Ill'snew work. The lower chamber was knownas St Triduana's aisle, containing relics of the saint, an altar dedicated to her, and

1Laing, David, Registrum Domus deSoltre . . . &c (Charters o f . . . Collegiate Churches in Mid-Lothian), Banna-tyne Club (1861), X L I V . 2 Register of th e Great Sea l of Scotland, n, N o. 1329.3 TheE x cheq uer Rolls of Scotland, ix, 540: '. . . Finlaio Fell, latumo, pro certis tegulis lapideis ad capellamdomini regis contiguam ecclesie parrochali de Lestalrig.' " Laing, op . cit., 274.6 R.C.A.M., An Inventory o f . . . the City of Edinb urgh (1951), 253. MacKinlay, James M., A nc ie nt ChurchDedications in Scotland: non scriptural dedications (1914), 477.6 Laing, op. cit., 280. ' Eyre Todd, George, TheBook of Glasgow Cathedral (1898), 317.


7/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA S AISLE 253appropriate fittings. It probably contained the saint's tomb-shrine also, though thisis hot stated. The upper chamber was a chapel of unknown dedication endowed byJames.The documents give no firm evidence for dating. James III came to the thronein 1460 at the age of nine, so it is unlikely that hisactive patronage of Restalrig - tobe associated with his unfortunate fondness for architecture - would begin until thefollowing decade. The upper chapel must have been well advanced by 1477 sincean altar existed in it at that date, though work was still being done on the roof in1486.

IV . PO ST- REF O RM A TI O N HIS T OR YOn aist December 1560 the General Assembly ordered that the 'kirk of Restalrig,

as a monument of idolatrie, be raysit and utterlie castin downe and destroyed'.1 Theparishioners were enjoined to repair to the church of South Leith, and within a fewyears Restalrig church was being robbed for building stone, which was used toremodel the Netherbow Port in 1571.2 In 1609, when South Leith was erected byAct of Parliament into the parish church, Restalrig is described as ruinous.3 Butfrom an unknown date the lower chamber of the hexagon was preserved through itsuse by the lairds of Restalrig as a burial vault.

The 1596 tombstone of Lady Jonet Ker (second wife of Sir Robert Logan ofRestalrig, 6th baron, concerned in the Gowrie conspiracy),4 which now lies withinthe lower chamber, is not conclusive evidence that she was buried there, for it mayhave been brought in at any time up to the late eighteenth century. William Burnettstated that the lower chamber was used as the family vault of the Logans of Restalrigfrom the sixteenth century, and that in 1608 the bones of the conspirator Sir RobertLogan were exhumed thence for their posthumous trial.5 I have not been able to findthe evidence fo r this statement. T he first unquestionable use for burial is in 1633when John Lord Balermino was 'buried in Restalrig's burial-place, being a vaultedaisle supported with pillars'.6

Arnot and Grose in the eighteenth century refer to the building as a burial-vault,without considering whether it may not originally have had another function.7 Bythis date the Earl of Moray was the proprietor: he purchased the estates in Restalrigand Leith, forfeited by Arthur Lord Balermino, when they were sold by the Baronsof Exchequer in 1755.Burials had continued until this date, filling the chamber to a depth of 6 or 7 ft. -about a foot below the capitals of shafts and central pier. When the building wasthus used the door in the NW. wall had been blocked, and a door had been formed1Booke of th e Universal! Kirke of Scotland, Bannatyne Club, i, 5.2 A diurnal o f Re m arkab le Occurrents, Bannatyne Club (1833), 241.3 Acts of the Parliaments of Scotland, iv, 442.4 Douglas, William, 'Armorial tombstone of Lady Jonet Ker at Restalrig, 1596' in P.S.A.S., L X I I(1927-8,) 27.5 'Restalrig and its collegiate church', in Trans . Scottish Ecclesiological Soc . , n, pt. 2 (1907-8), 274.6 Scot, Sir John of Scotstarvet, TheStaggering Sta te of theScots Statesmen ...(1754), 63.7 Arnot, Hugo, History of Edinburgh (1779), 256-7; Grose, Francis, Th e Antiquities of Scotland(1789-91), 1,43.8 Laing, op. cit., Ixi.


8/26


9/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA's AISLE 255And he continues:

The sacred usefulness of the well is surely testified by the fact that so late as 1840 its waterswere carried off in barrels by an ancient tenant. . ., and sold by him in Leith to clarify the visionof the dwellers in that enlightened burgh!

1

I do not know whether this is the memorable instance referred to by the Gildsman.Wilson and Burnett are the first authors I have been able to find who associate

the name of St Triduana with a holy well. And when the association is made,reasons are given for making it with St Margaret's Well, not with the Restalrighexagon.

V . R E S T O R A T I O NAfter a visit to Restalrig by the Scottish Ecclesiological Society in October 1906(the visit which occasioned William B urnett's paper referred to above) the sorry

state of the hexagon was commented on, and a committee was appointed to presenta memorial to the Earl of Moray, the proprietor. Moray directed that the buildingbe put in a state of repair and safety, and work began in August 1907 under thesupervision of Dr Thomas Ross.2

Dr Ross has given a description of the state of the hexagon before work began.[It] was filled with burials and earth to a depth of about six or seven feet. And on top of the

vault there was a grassy mound of earth about ten or twelve feet high, covered with bushes and treesof considerable size. This, along with a vigorous growth of bourtree in the adjoining open vaults,kept the place damp and dank, with a constant oozing of water through the chapter-house roof.The windows had been built up and portions of their tracery destroyed, the east one completelyso, and the sill cut away so as to form a wide and high doorway; the angle buttresses were cleanshaved away, and altogether the place was in a most deplorable and neglected condition.

At this time Dr Ross was convinced that the hexagon was a chapter-house. Hecontinued in this belief throughout the whole course of the works of restoration, andhis mind was only changed by a sequence of events at its close.

The progress of the operations begun in 1907 may be approximately recon-structed.The accumulation of earth in the interior was removed. As this was being done

water impeded progress; it was pumped out,and pumping was continued throughoutthe succeeding works.

While the works of restoration were going on the existing SE. door (an originalwindow opened out presumably after the Reformation as a door) was used for access.On their completion it was restored as a window and tracery inserted. The NW.doorway, hitherto unsuspected, was discovered in the course of the restoration. Theinternal and external rybats had been removed when the door was blocked, but fivesteps, and the facing of the ingoes, were found obliquely set in the thickness of thewall. This door was reinstated as the entrance.

1Trans. Scot. Ecclesiological Soc., n, 2 (1907-8), 278.2 Dr Ross's report on the restoration, illustrated by D. L. MacGibbon, is published in Trans. Scot. Eccle-siological Soc. , in, pt. 2 (1910-11), 238-46. This is the source for the following account.


10/26

256 PROCEEDINGS OF THE SOCIETY, 1963-64The blockings of the S. and SW. windows were removed and those parts of the

tracery which had been destroyed were restored.Some of the voussoirs of the inner arch of the large N. recess were missing and

were restored. So were the whole of the rear-arch and the masonry of the panelbetween the two arches. The back of this recess is now filled by a 1907 brick wall.The whole of the stone bench which runs right round the hexagon is a restoration.Dr Ross does not say how much, if any, of the original masonry existed in 1907. Butcomparison of the height of the bases of the attached shafts with the height of thebase of the central pier implies the existence of a bench or some other horizontalfeature at this level. The whole of the floor is also a restoration.

Only the NE. buttress is original. Dr R oss built new buttresses on three moreangles of the hexagon. The other two angles were obstructed by modern burialenclosures, so they were left unaltered.

On removing the high overgrown mound of earth above the lower chambervarious structural remains for an upper storey were found in position, and fragmentsof tracery, vault-ribs and six roof bosses were found lying at the bottom of the moundof earth. These structural remains and fragments have already been described. Theupper chamber was not restored. The rebuilt buttresses were surmounted by flatcaps, the wallheads levelled and a parapet built on their outer face. A pyramidalslated roof was constructed over the hexagon on a wooden frame.

It was decided permanently to exclude the water which was seeping into thecleared lower chamber, by tanking the interior below ground level with coats ofasphalt. Bo t h the exterior and interior faces of the walls were thus sealed. Inside, thewall-face was cut back about 6 in. A 4^-in. brick face was then built up to the lineof the original wall-face, and the cavity between the back of the brickwork and thecut-back medieval walling filled with asphalt ( P I . XL, 2).

The lower two-thirds of a rectangular aumbry (blocked at an earlier date, per-haps when the lower vault was first ,used for burials) in the NE. wall was cut awayas part of the preparation for the tanking.

The tanking of the floor was left until the last. I quote Dr Ross's account of thisoperation:

Nothing had as yet occurred to disturb the opinion that th e existing building was the chapter-house. A further excavation was made in the natural soil, when a bed of concrete was laid down,through which, however, the water bubbled up here and there in thin jets. A thick coat of asphaltwas spread over th e floor, but was completely destroyed by the inflow of water. M r Morgan, thebuilder, undertook to superintend th e relaying of the asphalt . B y dint of plugging th e jets and othermeans he succeeded and thereafter laid down thick stone pavement, which was further loaded withheavy stones laid on planks. The inflow of water was now stopped at all points except at the sumph,a hole sunk in the floor from which pumping was carried on. This was dried and filled with a bagof prepared concrete, when an iron plate was placed on the top. On the instant, with lightningrapidity, the plate and men on it were thrown to right and left, the water squirting out with violentforce in all directions, scattering the concrete like hail. This commotion lasted for a second or so,when the water was seen to ooze gently in the sumph as if nothing had happened. Afterwards a wellwas du g outside deeper than th e floor, and was built round in the hope that this would draw of f th ewater, but the same result followed a second trial. I t now became evident that th e building neverwas the chapter-house but a well - the long lost well of St Triduana!


11/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA's AISLE 257W e know from other sources that th e water reached at first a depth of 2 ft.

Thereafter the water-level fell, and seldom exceeded a depth of i ft . , while it fre-quently disappeared altogether after a period of dry weather. The fall in the water-level was attributed to alterations to a drainage scheme in the neighbourhood.1 Itwill appear from evidence yet to be given that the well-house interpretation did notoriginate with D r R oss . B u t w e m u st first consider the significance of the eventswhich have just been described.

A head of 2 ft. of water would produce a pressure of 124 lb. per sq. ft. Thispressure would be sufficient to force th e sealing out of position, though it couldhardly produce th e effect of extreme violence described. I n this connection it doesnot matter whether the pressure is being exerted by a spring rising within thehexagon, or by the general water-table throughout the immediate area. Digging awell outside would not affect the result in either case.

When the site was examined in 1956-7 the water did not flow into the excava-tions in the floor with the vigour of a spring: it seeped through all the sides of theexcavation, including that adjacent to the SW. wall, with a sluggish persistence.

The evidence which seemed to confirm the well-house interpretation for Dr Rosswas provided by M a jo r George Logan Home of Edrom. In his History of th e LoganF a m i l y , published in 1934, he deals at some length with the Restalrig hexagon, andat one point with its restoration:

In 1907, by direction of the Earl of Moray, the crypt was restored and the earth and rubbishremoved. The architect, who did not know it was St Triduana's Well, had great difficulty in fightingthe constantly rising water. . , . Nothing would keep it down, and the story of St Triduana havingbeen brought to his notice, the architect realised that it was the crypt and well of the saint. . . .When finished, clear water rose to a height of about two feet in the well, which covers the whole floor,with the exception of a stone gallery round the sides. Thus the victory rests with the saint. . . .2

Two feet of water would in fact comfortably cover the stone bench, which is 9 in.high.In a letter printed in the Scotsman on 5th December 1931, Major Logan Home

makes it clear that he had himself told Dr R oss of the legend: 'I then brought theold legend of St Triduana to the notice of the architect.' The version of the legendwhich Logan Home would tell Dr Ross appears in another place in the History of th eL o g a n F a m i l y . Logan Home begins with a translation of the legend from the A be rde e nBreviary; and continues by naming Camerarius and Hunter the Dominican asauthorities for St Triduana in Scotland.

Camerarius also gives the legend of St Triduana, and Hunter the Dominican, testifies thatthis most holy virgin was in great honour among the Scots for the sanctity of her life and the gloryof her miracles, and that numerous chapels were built in her honour . . . she was buried at Restalrig,and when her body was committed to the tomb, a well of pure water sprang up from the groundwhich has ever since been renowned for its healing virtues to the eyes.3

The passage about the well of pure water is neither in the A be rde e n Breviary nor inChambers (Camerarius). The context seems to imply that it occurs in the work of1Douglas, W., in P.S.A.S., LXII (1927-8), 34. 2 History . ..,28. 3 History . ..,25-26.


12/26

258 PROCEEDINGS OF THE SOCIETY, 1963-64Hunter the Dominican. But this work was already lost by the nineteenth century.Logan Home gives no other references.

Yet it is obvious that this version, presented to Dr Ross at the disturbing pointwhen the supposed chapter-house was flooded, would provide a most attractivesolution to the new problem. Dr Ross readily accepted it as the true solution.Thenceforward the 'chapter-house' became 'St Triduana's Well-house', and thenew interpretation was accepted without question by all subsequent writers onRestalrig.1

V I. WELL-HOUSESThe closest parallel to the Restalrig hexagon is St Margaret's Well, a building

similar in plan, and having a vault which in layout is identical with that of the lowerchamber. There are no other examples of this type of vault in Scotland.

This resemblance alone suggests that one hexagon is a copy of the other; andthere are several indications (the non-functional reproduction at St Margaret's ofthe bench at Restalrig, and the clumsiness of the compressed rib-pattern of theSt Margaret's vault), that St Margaret's is a miniature, simplified reproduction ofthe lower chamber at Restalrig. The little building of St Margaret's is undoubtedlya well-house. Is this not a strong argument that its prototype had the same function?B ut there are two important differences between the hexagons. The first is thescale (fig. 3a, d). The area of the Restalrig hexagon is more than ten times greaterthan that of St Margaret's Well, which is a simple vaulted cistern. Second, in StMargaret's the water flowed out: of the lower part of the structure and thence outsideover the sill. It is not clear whether the water rose from the floor as a spring, or fellfrom a tank above the vault through the mask protruding in relief from the centralpillar. The drawing in Billings,2 made before the removal of the building to theQueen's Park, does not show water issuing from the mask, though it does issue fromit now by way of a modern spout in the mouth of the mask. But there is no escapefor the water from the structure at Restalrig. Once inside it would stagnate unlessit rose high enough to flow over the sills of the windows at the level of the churchyard.

The lack of an outflow is only one of several problems which at once arise whenthe use of the lower chamber as a well-house is considered.

If the water rose above a height of 3 in. it would cover the moulding of the baseof the central pier, which seems to be meant for display. (St Margaret's Well has nomoulding on the base of its central pier.) At any depth water would cover the wholeof the interior area except the recess in the N. wall and the stone bench round thesides (unless it rose to a height of 9 in., which it frequently did, when the stone benchwould be covered too. The outflow from St Margaret's Well is below the level of itsbench.) The stone bench is too narrow to be anything but a bench. It is not a

1Michael Barrett may be quoted. In his Ca l enda r of Scott ish Saints, published in 1904: '[St Triduana's]tomb became a favourite place of pilgrimage. Before the Reformation it was the most important of the holyshrines near Edinburgh'. Barrett follows the authority of the Aberdeen Breviary for the tomb. In his Footprintsof theAncient Scottish Church published, after the restoration, in 1914: '. . . St Triduana, a virgin recluse ofabout the eighth century, was enshrined [at Restalrig], and a holy well, bearing her name, was renownedfor the cure of diseases of the eye.'z Billings, Robert, Th e Baronial a nd Ecclesiastical Antiquities of Scotland, 1845-52, n, PI. 30 .


13/26

1 8 3 6 W E S T E N DO F C H U R C H

B A S E O F A T T A C H E D S H A F T

!83b W E S T E N DOF CHURCHLA N OF L O W E R CHA M B E R O F H E X A G O N

AND FLOOR LEVEL OF C H U R C H1907-8 T I M B E R

ROOF A N D P A R A P E T

S E C T I O N A T A - A

PLAN OF R E M A I N S OF UPPERCHA M B E R O F H E X A G O N

S C A L E O F F E E TF I G . i. Restalrig church and hexagon: present state [to face p. 258


14/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA's AISLE 259gangway. The steps from the door into the lower chamber lead immediately to thepool of water.

The force of these objections is emphasised when an extant well-house of similarsize is considered. Comparative material for well-houses is scanty: there is nothingrelevant in Scotland; English examples are simple and small. I have found no usefulcontinental analogies. But at Holywell, in N. Wales, there is a building with ele-ments strikingly similar to those of Restalrig: St Winifrede's Well is a two-storeyedrectangular structure of which the lower storey, centrally planned, is a well-house;and the upper storey is a chapel. The internal area is equal to that of the Restalrighexagon.1

The planning of St Winifrede's is strikingly different from Restalrig's. The well-house is laid out as a functional unit. The spring itself rises into a star-shaped tank,about 9 ft. across. Loose cobbles lie on the bottom of the tank, and the water risesvigorously inside. On the E. side of the tank is a rectangular bath paved with stone,entered by a flight of steps at one end,and left by a second flight at the opposite end.The water flows from the tank into the bath, out of the well-house into a largerectangular bathing-pool outside and finally down a steep slope.

The tank and bath together occupy only about one-sixth of the internal area ofthe well-house. The rest of the area is taken up by an ambulatory surrounding thetank and bath. It is the ambulatory which is first entered from the door.

Near St Asaph another well-house, St Mary's Well, repeats the planning patternof star-shaped tank, bath, and ambulatory, though here the chapel is on the samelevel as the well-house.In contrast to the Welsh well-houses, there is no evidence of arrangements atRestalrig to deal with the circulation of either water or pilgrims.These objections are not of course conclusive. The Restalrig hexagon may have

been a badly planned well-house. A fter all,in modern times it often containedwater. But the presence of the water may be otherwise explained. Either the watertable may have risen in modern times with the development of the Restalrig area,or the water may have been unexpected when the chapel was begun. It has beenobserved that the water-level fluctuates: sometimes it is several inches below floorlevel. The subsoil is gravel, not apparently waterlogged. If the foundation of thehexagon and its central pillar were laid down in a dry season, work might havecontinued for some time before any intrusion of water was observed. When waterwas found to intrude, the floor would be established at a higher level-subsequentlydisturbed by the post-Reformation burials. Such a blunder is not unknown inmodern building experience, though it would be improper to quote parallels.

VII. A N A L O G I E SThe argument so far suggests that the Restalrig hexagon is a fifteenth-century

two-storied chapel. I have made an unsuccessful search for parallels in buildings in1Royal Commission on the Ancient andHistorical Monuments ... in Wales . . .,n, County of Flint (1913),No. 127. Plan reproduced as fig.y by permission.


15/26

260 PROCEEDINGS OF THE SOCIETY, 1963-64Great Britain and on the Continent; a true parallel should correspond to the hexagonboth in design and in function. I summarise the background of the elements ofRestalrig's planning below.

There are three main types of two-storied consecrated buildings.In the first, the building is centrally planned, and its two stories are united byan open central area, as in the Doppelkapel lm in Germany.1

The type stems from the Imperial Chapel at Aachen, which is in turn inspiredby San Vitale, Ravenna.2 At Aachen the upper storey is a spacious triforium orgallery for the imperial household. There are no comparable buildings in Scotland;the square Bishop's Chapel at Hereford of 1079-95 is the only example in GreatBritain.3

This type of two-storied chapel is clearly not relevant to Restalrig and need notbe considered further.

The second type, with separate stories, is only occasionally centrally planned.Buildings of this type have an altar in both of their stories. Bandmann4 traces itsorigin from Early Christian mausolea to the two-storied rectangular Camera Santaat Oviedo consecrated in A .D . 802. The latter had two separate stories of which thelower contained the body of the saint and the upper the shrine.5 The most out-standing example of the type is la Sainte Chapelle in Paris, built in 1246-8 as areliquary chapel. La Sainte Chapelle and its imitations are rectangular. Andalthough there are two-storied polygonal examples the type is very rare after thethirteenth century in Europe.

In Scotland there are some late medieval two-storied ecclesiastical buildings inthe form of transeptal aisles. The Blacader aisle adjoining Glasgow Cathedral is thelower storey of a late fifteenth-century chapel intended to be of two stories but nevercompleted. The floor of the lower chamber of the aisle is sunk below the floor-levelof the lower church - for the nave of the cathedral is itself of two stories. TheBlacader aisle is relevant here in that it shows a contemporary interest in two-storiedplanning in Scotland.

The third type is a two-storied building with the lower chamber used as a burialvault (usually for family burials, though sometimes as a common charnel-house orossuary) the upper usually containing holy relics and being used as a chantry.6There are several continental examples of the type ranging from the twelfth to thefifteenth century, most being centrally planned.7 In this type there is, however, no

1Bandmann, Gunter, s.v. 'Doppelkapelle, kirche', in Reallexikon zur deutschen Kunstgeschichte, edd.Schmitt, O., Gall, E. and Heydenreich, L. H., Stuttgart (1933-) lief, ir, col. 196 ff. There is no agreeddefinition of Doppelkapelle. For Bandmann it is simply a two-storied chapel: it may be of two stories con-joined by an open central area (Type A) or two separate stories (Type B). But other authorities wouldexclude the second type from the definition, e.g. Schurer, Oskar, Die Kaiserpfalz Eger , Berlin (1934), 92.2 The ultimate derivation is the Constantinian Holy Sepulchre rotunda in Jerusalem.3 Drinkwater, Norman, 'Hereford Cathedral, Bishop's Chapel of St Katharine and St Mary Magdalene',in A r c h . Journal (1954). 4 op. cit., col. 20 1 ff . : Bandmann's Type B.5 S ee Grabar, Andr, Martyrium, Recherches sur le culle de s reliques et t'art chretien antique (1946), I, 574-5.6 Bandmann, op. cit., col.197. Buildings of this type are not Doppelkapellen by any definition.7 e.g.the probably twelfth century 'grand Octagone' at Montmorillon, Vienne: see Charles Grosset inCongres archaeologique d e France (1951), 199-206; and an example of 1440 in Brandenburg, noted by HeinzBiehn, Ein Beitrag zur Geschichte de s deutschen ^entralbaues b is zum Jahre 1500 (1933), 76.


16/26

LAN OF LOWER CH A P E L A N DLOOR L E V E L OF C H U R C H

C O N T E M P O R A R Y V A U L T O F L O W E R C H A P E LV \\ O M I T T E D FOR C L A R I T Yv>

L_J

10 5 OH-M i I i ' i i I

IO 20 3O 40f

5O

S C A L E O F F E E T

K E YE X T A N T fro

$$$81 clrc

a b l y e a r l y 1 5 t h . c e n t u r y

1470-86

r I486

P L A N OF UP 'E R C H A P E LAT F L O O R L EVELn ..; x n r [ 7

F I G . 2. Restalrig church and hexagon: medieval structures

,VL I N E OF R O B B E D\ A S H L A R F A C E

{to face p. s6o


17/26

KING'S CHAPEL , RESTALRIG AND ST TRIDUANA'S AISLE 261functional parallel to Restalrig, where the employment of the lower chamber forburials represents a secondary use.

These three types are consecrated buildings. There also exists a number of two-storied centrally planned ancillary ecclesiastical buildings, of which the best knownare chapter-houses.1 Most of the chapter-houses are thirteenth-century octagons(none is hexagonal), and most have a single storey, though Lichfield, Westminsterand Wells have two stories, of which the lower storey is sometimes sunk below groundlevel as at Westminster. The series continues into the fifteenth century with single-storied examples at Manchester, Warwick and Elgin, where the rebuilding of anearlier vaulted octagon in the late fifteenth century provides the latest example in theScottish series. There are also octagons at Glenluce and Inchcolm (thirteenth cen-tury) and the foundations of another at Holyrood (c. 1400). The nineteenth-centuryassumption that the Restalrig hexagon was a chapter-house was thus well-foundedon comparative material.

Chapter-houses are found in England at Manchester and Warwick attached tocollegiate churches. But no Scottish collegiate church has a chapter-house, thoughseveral have sacristies. During the period when the Restalrig hexagon was believedto be a chapter-house no one commented on its size: it is substantially larger thanall the Scottish chapter-houses attached to cathedrals or monastic houses with theexception of Elgin.

Though the chapter-houses are the best known, they are not the only Britishexamples of late-medieval polygonal planning. Cloister lavatories take this form,and there is a hexagonal example at Sherborne, Dorset, built 1504-35.2

None of the vaulted examples of centrally planned buildings which I have foundhave the peculiar design of the lower vault at Restalrig. By far the greater numberhave variants of the simple plan of radial ribs. The few vaults which resembleRestalrig in plan have no central pillar and are thus not true parallels.3

So far the search for source-parallels has been confined to architecture, and it hasbeen unsuccessful. But there are many late medieval metal objects - reliquaries,censers, ciboria - which have a closer resemblance to the Restalrig hexagon than anybuilding I have been able to find.

The upper part of a German sixteenth-century gilt bronze reliquary in the Burrellcollection is hexagonal with large traceried windows and flying buttresses springingfrom the base of a pyramidal roof which has crocketed and pinnacled dormers. The

1A succinct descriptive list is given by Bilston, John, 'On the discovery of some remains of the chapter-house of Beverley Minster', in Archaeologia , L I V , pt. 2 (1895), 425-32.2 Godfrey, Walter, 'English Cloister Lavatories', in A r c h . Journal, cvi supplement (1952).3In the vault of the Abbot's Kitchen at Durham the ribs of the octagonal chamber are thrown trans-versely across the compartment in the same manner as at Restalrig, though at Durham the reason is to leavea central void to act as a chimney. The same principle, with the same object, is seen in the twelfth centuryoctagonal abbey kitchen at Fontevrault, Maine-et-Loire, where the whole roof is supported by a series ofsquinch arches (Congris archaeologiqite de France (1910), 58-60). A closer resemblance appears in the late-thirteenth-century N. porch of St Mary R edcliffe, B ristol. Here a most ornate hexagon of a single loftystorey has a vault which is constructed on exactly the same rib-plan as lower Restalrig. But the absence ofa central pillar again makes it valueless as a structural parallel, however close the similarity in design.


18/26

262 PROCEEDINGS OF THE SOCIETY, 1963-64relic was held in a cylindrical glass under thehexagon. Two Italian fifteenth-centurysilver censers in the Royal Scottish Museum1 resemble two-storied hexagonal build-ings with large openings and pyramidal roofs. The principal part of a Spanishfifteenth-century silver gilt ciborium in the same museum is a buttressed hexagonalbuilding of one storey with a pyramidal roof.2 The three fifteenth-century StAndrews University maces and the Glasgow University mace are crowned byhexagonal tabernacles.3

These metalwork objects are all variations of a single- or two-storied hexagon.B ut the variations do not seem to be inspired by real buildings: the hexagon is oneof the rarest plans, the multi-storied hexagon rarer still. Had the craftsmen beeninspired by anything they had seen in stone and lime, an octagonal plan would havebeen a more natural choice. These metalworkers were no more imitative than weremedieval manuscript illuminators in many of their representations of buildings. Themetal objects are an architectural fantasy.

Restalrig's inspiration may not be architectural. The hexagon may be modelledon a comparatively small metal or wooden prototype, which could have been areliquary or even the ornamental shrine of Triduana.4 The suggestion can be nomore than tentative, though it would assort well with the enterprising architecturalenthusiasm of Restalrig's patron, James III,who 'delytit mair in musik and polliecieof beging [building] nor he did in the government of his realme'.5

VIII. C O N C L U S I O N SThe conclusions of this investigation may now be summarised.There is no sound reason to believe that the lower chamber of the Restalrig

hexagon was a well-house. Objections to this interpretation have been discussed indetail.The combined evidence of the structure and of medieval documents suggest that

the hexagon was a two-storied chapel built by James III and known as the King'sChapel. The upper chamber contained a chapel of unknown dedication. The lowerchamber, known as St Triduana's A isle, contained the altar and relics, perhaps alsothe tomb, of the saint.The building of the King's Chapel may have commenced at any time in theperiod 1460-77, though in view of the tender age ofJames III when he came to thethrone, it is more likely in the later part of the reign. The structure was largelycomplete and functioning by 1477.

When the hexagon was built (fig. 2), it adjoined the present parish church, asimple structure probably of the early fifteenth century, with unicameral chancel and1 R.S.M. (1892), 300, 301.2 R.S.M. (1902), 217. Besides the examples noted, there are five other 'two-storied' hexagonal turriformcensers (1892: 326, 342; 1896: 329, 330, 343) and a reliquary with a hexagonal turriform cap (1892: 316)in the museum.3 Brook, Alexander J. S., 'An account of the maces of the universities of St Andrews, Glasgow, . . .' inP.S.A.S., xxvi (1891-2), 440-82.4 Shrines and monstrances could be turriform as well as the smaller objects noted above.5 Lindesay of Pitscottie, Robert, The History andCronides of Sc ot land, ed. Aeneas J. G. Mackay (1899-1911), I, 163.


19/26

B

LOWER VAULT U P P E R VA UL TST. W I NIFRED'S W E L L H0L O W E R S T O R E Y ( C H A P E L

ST. M A R GA R E T' S WE L L RILSTALRIG(NOW IN QUEEN'S P A R K )

P L A N S A N D I S O M E T R I C S K E T C H E S O F V A U L T Siafframs of hexaeon vaults: comnarisnn of hexagon to St Winifred's Wpl] and St Mnrffarpt's

FEET O IO 20J_T R I C S K E T C H E S O F V A U L T S l i t . . M , nt_______I-

F I G . 3. Diagrams of hexagon vaults: comparison of hexagon to St Winifred's Well and St Margaret's Well. (Plan of St Winifred's Well reproducedon the Ancient Monuments of Wales)

3Oby per

LY W ELLA B O V E ;

4O. I .. soLsion of the Royal Commission[t o face p. 2 62


20/26

KING'S CHAPEL, RESTALRIG AND ST TRIDUANA's AISLE 263nave and a N. sacristy. An ambitious rebuilding of the church with a chancel andaisled nave was commenced and very shortly abandoned. The commencement mayhave coincided with the completion of the hexagon; the abandonment with thedeath ofJames III in 1488 and the consequent loss of personal royal interest.

The buildings of the collegiate church may have been more extensive, but noneare known apart from the hexagon, the fifteenth-century church with N. sacristy,and the unfinished rebuilding of the church.

St Margaret's Well may have been associated with Restalrig church and mayhave been the source of revenue in the garden of the college noted as one of theperquisites of St Triduana's prebendary. St Margaret's is copied from the interiorof the King's Chapel.

The King's Chapel is not copied from anything. Its design consciously or un-consciously unites in one building a large number of elements found widely separatedin time and place elsewhere. It may have been immediately inspired not by abuilding but by a ritual object.A s designed and built the chapel would be a much more striking work than thesurviving remains immediately suggest. In external elevation, the sunken lowerchapel is merely the substructure for the much more spacious upper chapel, whosewindows were probably a prominent feature of the facades. The detail of the win-dows cannot be recovered but their large size is clear from the surviving fragments.It may be conjectured that the upper wall-faces and wall-heads would be decor-ated to contrast with the plain surfaces of the lower chapel. The walls wouldprobably be finished with a parapet to conceal the roof, with the buttresses carriedup into pinnacles.The unknown master-mason who designed the Restalrig hexagon created animaginative and original composition in the ornate but unadventurous context offifteenth-century Scottish ecclesiastical architecture.

Th e Society is indebted to the Ministry of Public Building andWorks for a grant towards the cost of this paper


21/26


22/26


23/26


24/26


25/26


26/26

Documents

The King's Chapel at Restalrig and St Triduana's Aisle